Re: DKRs without the "K"

From: Lee Iverson (
Date: Wed Feb 14 2001 - 08:01:31 PST

In message <>, Eugene E
ric Kim writes:
>Groves are directed graphs of nodes, so the data structures you propose
>above qualify as an implementation of a grove. A property set describes
>how the grove is constructed and what it consists of. If you want a
>DOM-like data model for manipulating a document, you create a property set
>that defines element nodes, attribute nodes, etc. If you want a data
>model like the one described above, you create a property set that defines
>text nodes and structure nodes and data nodes.
>I don't think groves conflict at all with what you're proposing. In fact,
>I think we're essentially thinking about the same picture (you can
>correct me if I'm being presumptuous), although I'm not sure how you would
>specify the data models with DTDs. Basically, you're proposing a grove
>constructor which has a property set parser. In order to construct a
>grove, you need a plug-in for that document type that consists of a
>document parser and a property set. The document parser creates a parse
>tree, then the grove constructor takes that tree and the property set to
>construct the grove.

After reading a few of the grove introductions, I'm way psyched. That is
exactly the conclusion I've come to. What I'm designing is a distributed
grove engine. I need to understand more about the granularity, but the idea
of having a few basic building blocks for all structured document types and
a means of manipulating them in a distributed environment is going to be
a real killer app.

>I think the twist that you've thrown in (which you first introduced when
>you suggested the DDOM many moons ago) is that your grove implementation
>is designed to work in a distributed environment, with features like
>version control. That's cool stuff.

Lee Iverson SRI International 333 Ravenswood Ave., Menlo Park CA 94025 (650) 859-3307

This archive was generated by hypermail 2.0.0 : Tue Aug 21 2001 - 17:58:01 PDT